Not available in English yet
给 AI 系统下毒,发现它比 2015 年微服务还脆弱 — 混沌工程开始被认真对待
Related Reading
More on #LLM
On-Device LLMs Are VRAM-Hungry — A Dev Open-Sources a Patch to Borrow Idle Home GPUs
NVIDIA wants AI workstations on desks, but local LLMs choke on VRAM. A Reddit dev open-sourced a patch borrowing idle home GPUs. Local isn't cheaper —
GPT-3 Retired Today — The Model That Taught the World 'AI Can Write' Ends an Era
OpenAI officially retired GPT-3, the original davinci model, this week—five years after its 2020 launch. Developers react with nostalgia and migration
AI Service Crashed 12% of Users for 28 Minutes — Engineering, Not the Model
AI inference outage postmortem: 12% of user requests timed out for 28 minutes. Root cause: missing queue length limits, not the model.
50 Records In, 20 Seen by Model: Agent Silent Truncation Is the Real Killer
A tech team found AI agents read 50 database records but silently passed only 20 to the model with no errors raised. Traditional "no error = success"
Jev packages itself as a new AI paradigm. Veterans say: it's just a classifier.
Reddit's top post: Jev's "System One Model" replicates pre-LLM zero-shot classification. BGE-small with classical stats hits 93.3% vs Jev's 83.2% on B
Enterprise AI Knowledge Bases Miss the Point — Bug Sits in Retrieval, Not LLM
RAG is now standard for enterprise knowledge bases but keeps misfiring. We trace the fault to retrieval, not the LLM. New tools mark its maturation.